|Publication number||US5490153 A|
|Application number||US 08/286,030|
|Publication date||Feb 6, 1996|
|Filing date||Aug 4, 1994|
|Priority date||Aug 4, 1994|
|Publication number||08286030, 286030, US 5490153 A, US 5490153A, US-A-5490153, US5490153 A, US5490153A|
|Inventors||Thomas A. Gregg, Joseph M. Hoke, Albert Ing, Chin Lee|
|Original Assignee||International Business Machines Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (2), Referenced by (57), Classifications (9), Legal Events (8)|
|External Links: USPTO, USPTO Assignment, Espacenet|
The present invention is generally directed to data communication systems using optical fibers to carry information. More particularly, the present invention is directed to recovering frames of information which have been corrupted during transmission but without incurring substantial time out penalties.
Fiber optics have enabled the transmission of long strings of data in a serial fashion from a driver to a receiver over long distances (typically measured in kilometers) at very high data rates (typically specified in billions of bits per second). This is in contrast to more traditional communication over electrical wires which only allow data to be transmitted for relatively short distances at these high data rates. The distances for communication over wire means is typically in the range of several tens of meters.
Fiber optic data transmission is however inherently noisy in that bit errors in the data are frequent. Error rates of one in a trillion or even one in a billion bits are common. Various checking methods including cyclic redundancy codes are used to detect these errors.
Several methods have been suggested for recovering data that has been transmitted but received with error indications. One method is to employ a high level protocol that keeps track of the time from request to response. Each time a request is sent from one side of the fiber optic link to the other, the sender starts a timer. If a response is not received within a specified time, the sender "assumes" that either the request or the response was lost. The sender then requests status from the other end of the link to determine if the request should be resent.
Another method of recovering data received in error is to package the data into frames and to assign a sequence number to each frame. If a receiver detects a frame with a sequence number that is out of order, it assumes that one or more frames were lost. Using the sequence number of the last correctly received frame, the receiver then requests the lost frames to be retransmitted. This method allows multiple requests and responses to be on the link at the same time, thus improving the utilization.
However, using timers to detect damaged data, as described above, is inefficient since the time period chosen has to be long enough to allow the slowest anticipated response. The problem with the other suggested solution using sequence numbers is that sequence numbers add complexity to the transmission and reception hardware.
Thus, what is needed is an improved method for determining which data frame is damaged by bit errors. This method should be fast and it should not disturb traffic on the same link due to other non-related operations typically present in a multiplexing environment.
The present invention is embodied in a system and method for asynchronously transmitting data blocks between two information handling systems. Two carriers are used to interconnect two systems and to provide serial data transmission in both directions. Together, these carriers and the supporting hardware is called a link. Information frames are provided as a mechanism to transmit data serially on the fiber. A frame contains all or part of the contents of: a request; a response; or a data area.
The protocol for data transmission includes a request sent from one information handling system to another, followed by optional data areas transmitted in either direction, followed by a response sent back to the requesting system. Each sequence of "request, data, and response" is called an operation, and multiple operations may be interleaved over the link. For each operation, strict ordering of the request, data, and response is maintained. The system maintains state information as to the progress of each operation that has originated or to which it is responding. To maintain the state information, the frame headers of the requests and responses include an indicator (the A bit) informing the receiver that data areas are associated with either the request or response.
According to one embodiment of the present invention, the receivers of the frames use receiver state information to detect the correctness of the received frames. In particular, in some cases when a response is received with the associated data indicator set, the receiver may determine that the data area was lost. The receiver may then instruct the other end of the link to retransmit the lost data area. Similarly, in some cases when a data area is received and was not preceded by a request, the receiver can assume that the request was lost and instruct the other end of the link to retransmit the request.
The subject matter which is regarded as the invention is particularly pointed out and distinctly claimed in the concluding portion of the specification. The invention, however, both as to organization and method of practice, together with the further objects and advantages thereof, may best be understood by reference to the following description taken in connection with the accompanying drawings in which:
FIG. 1 is a block diagram illustrating a physical link between two information handling elements;
FIG. 2 is a block diagram illustrating a multimessage channel buffer structure;
FIG. 3 illustrates the format of an exemplary frame, particularly showing the presence of a link control word;
FIG. 4 illustrates the contents of the Link Control word;
FIG. 5 illustrates sequences of command, data and response exchanged between two computing elements or systems;
FIG. 6 illustrates additional sequences of command, data and response exchanges where multiple data transfers are performed;
FIG. 7 illustrates an exchange sequence where one of the frames has a transmission error;
FIG. 8 illustrates the frame reception state table for the originator of an operation;
FIG. 9 is a list of originator actions;
FIG. 10 is a list of originator events;
FIG. 11 is a list of originator states;
FIG. 12 illustrates the frame reception state table for the recipient of an operation;
FIG. 13 is a list of recipient actions;
FIG. 14 is a list of recipient events;
FIG. 15 is a list of recipient states.
Turning first to FIG. 1, a physical link between two computing elements 102, 104 is illustrated. These elements could be, for example, two computers or a computer and a shared memory device or in general, any information handling system. In any event, computing elements 102, 104 are connected by way of intersystem channel link 106 including fiber optic cable link 108. Fiber optic link 108 is formed with optical fiber cable pair 110. A fiber cable pair includes two optical fibers, one for transmitting information and one for receiving information. Fiber cable pair 110 is coupled to computing elements 102, 104 by way of transceivers (TCVR) 112 and 114 located at opposite ends of the link. Each of the transceivers 112 and 114 includes a transmitter unit and a receiver unit.
All of the data traffic over fiber optic cable link 108 supports message passing between the computing elements 102 and 104. A typical message is a request sent from computing element 102 to computing element 104. Data may be associated with the request and is either sent from computing element 102 to computing element 104 (a write operation) or from computing element 104 to computing element 102 (a read operation). After the data is transferred, a response is sent from computing element 104 to computing element 102. The messages, consisting of requests, data and responses are stored in buffers located in both computing elements as shown. To transfer a request, data, or response, transmit buffers 116 and 122 and receive buffers 118 and 120 are employed in computing elements 102 and 104. It should be understood that transmitting buffers 116 and 122 may be located anywhere in the transmitting computing elements 102 and 104, including in main processor storage. It should be further understood that the receive buffers 118 and 120 should at all times be immediately accessible by transceivers 112 and 114. Therefore, receive buffers 118 and 120 are usually implemented as arrays dedicated to the link, and they are not in main processor storage where access is shared among many different elements within computing elements 102 and 104.
To process a complete message with data requires buffers in both computing elements 102, 104. The computing element that initiates the message is the originator, and the computing element that processes the message is the recipient. FIG. 2 illustrates a typical situation in which there are multiple buffers on both sides of a link. For example, transmission of a message with data from Channel A (202) to Channel B (204) requires Channel A originator buffers shown in block 206 and the Channel B recipient buffers shown in block 208. Each group of buffers, 206 and 208, are called "buffer sets". When a message is sent, originator buffer request area 210 is loaded with the request, and the request is sent over the link to recipient buffer request area 216. If data is to be transferred, it is either sent from originator buffer data area 214 to recipient buffer data area 220 for write operation, or it is sent from recipient buffer data area 220 to originator buffer data area 214 for a read operation. After data transfer, if any, the response is loaded into the recipient buffer response area 218 and sent across the link to originator buffer response area 212 or vice versa depending on the information flow direction.
The information that is transferred from one side of the link to the other side is contained in frames. This information is always targeted to a particular buffer area, and the targeting information is contained in the link-control word (see below) of the frame. This targeting information allows the frames to be multiplexed over the link in any order. As an example, returning to FIG. 2, Channel A could send a request for buffer set 1 followed by write data for buffer set 0, followed by a response for buffer set 0, etc. It should be understood that a computing element can dynamically set up various numbers of originator and/or recipient buffers depending on the number and type of links to be established.
The format of an exemplary frame is illustrated in FIG. 3. When no frames are being transmitted, idle words 310 are continuously sent on the link. When frames are transmitted, they start with a data word which is the link-control word (LC) 302. Various fields in the link-control word 302 identify the frame format and type, designate a buffer set area, and control the state of the transceiver and link. These fields are described in detail below.
A link-control-CRC (cyclical redundancy check) word 304 preferably follows link-control word 302. Link-control-CRC word 304 is conventionally generated from values in the link-control word. Link-control-CRC word 304 is checked at the receiver to test the validity of the link-control word in the incoming frame.
There are two types of frames: control frames and information frames. Control frames do not have an information field. Control frames consist only of a link-control word and a link-control-CRC word. An information frame has link-control word 302, a link-control-CRC word 304 and an information field 306. Information field 306 contains, for example, from one to 1,024 words. Information field 306 contains the information sent from a buffer set area at one end of the link to a buffer set area at the other end of the link.
An information field is followed by information-field-CRC word 308. Information-field-CRC word 308 is conventionally generated from the values in the information field. Information-field-CRC word 308 is checked at the receiver to test the validity of the information field in the incoming frame.
FIG. 4 shows details of link-control word 302. Format bit 402 indicates whether or not the frame contains an information field. Requests, responses, and data frames all have an information field while acknowledgements and rejects do not have an information field. Type field 404 specifies that the frame is either a request, a response, or a data frame. Buffer set number 406 specifies which buffer set is the target. "A bit" 408 has two uses. In a request frame, "A bit" 408 is used to indicate that data frame(s) are to follow (a write operation), and in a response frame, the "A bit" is used to indicate that data frame(s) preceded the response (a read operation). In a data frame, the "A bit" is used to indicate that more data frame(s) are to follow.
It should be kept in mind that, as used herein, the terms "data frame" and "information frame" are not synonymous. Requests, responses and data frames are all information frames (that is, frames with an information field). A request is called a Message Command Block (MCB) and a response is called a Message Response Block (MRB) (see FIGS. 5 and 6).
Information transferred to a particular buffer may be contained in more than one frame or frame group. The first frame for a buffer area always has Start 410 bit "ON" and this bit also indicates the validity of Block Count 412. This count indicates the total number of 256 byte blocks that are transferred to the buffer. This count does not indicate the length of the presently transmitted frame. The transmitter can end the frame with a CRC 308 word on any 256 byte boundary of the information 306 field. When the transmitter resumes the transfer to the buffer, it starts the new frame with Start bit 410 in the link-control word reset to zero. The zero value of the start bit indicates that this frame is a continuation of a previous frame targeted to the same buffer. The receiver knows that all of the information has been received when the total number of 256 byte blocks transmitted, in all of the frames, have been received and satisfy Block Count 412, transmitted in the link-control word of the first frame. A buffer area can be transmitted by any number of frames from one to the total number of 256 byte blocks. For example, a 1024 byte buffer area can be transmitted in from one to four frames.
Although not provided with sequence numbers, each frame is interlocked with subsequent and previous frames. The interlocking is accomplished by the protocol on the link. For example, each message for a particular buffer set starts with a request, is followed by data and is subsequently followed by a response. Each of these types of transmissions has a unique link-control word since each transmission is targeted to differing buffer set areas. FIGS. 5 and 6 illustrate the protocols.
FIG. 5 shows three operational examples. The first example shows a request and response with no data transfer. In this example, originator 502 sends a request in Message Command Block (MCB) (in step 506). The LC 302 for this frame has "A bit" 408 set to zero (indicated by the bar over the "A" in the Figure) since there is no data to follow. After the request has been processed, recipient 504 sends a response in Message Response Block (MRB)(508). The LC word for this frame also has the "A bit" set to zero since there was no data preceding the response.
The second example in FIG. 5 is a write of a single data area. In this case, MCB (510) has the "A bit" set to "1" since there is at least one data area to follow. After the MCB, the originator sends the data area DATA (512). The "A bit" in this DATA frame is set to zero because there are no more data areas to follow. After recipient 504 processes the request and its associated data, it sends a response, in the form of MRB (514). The "A bit" in the MRB is set to "0" since there was no data preceding the response.
The third and last example in FIG. 5 is a read of a single data area. In this case, MCB (516) has the "A bit" set to zero since there are no data areas to follow. Recipient 504 processes the request and returns data area DATA (518). The "A bit" in this DATA frame is set to "0" because there are no more data areas to follow. After recipient 504 sends the DATA frame, it sends response MRB (520). The "A bit" in this MRB is set to "1" since there was at least one data area preceding the response.
FIG. 6 shows two examples of transferring multiple data areas. In the first example, a write operation transferring two areas is performed by the originator. As in the single data area example, MCB 602 and first data area 604 are sent by the originator. First data area 604 has the "A bit" set to "1" indicating that more data areas are to follow. The recipient processes the first data area by moving it to main storage (or elsewhere) thus freeing the buffer area for the receipt of the next data area. Next, the recipient sends an acknowledge ACK (606) frame. This frame contains no information field but the link control word identifies the buffer set. The originator responds to the ACK frame by sending the next (and last) data area DATA (608). The "A bit" in this DATA frame is set to "0" because there are no more data areas to follow. After the recipient processes the request and its associated data, it sends a response, an MRB 610. The MRB has the "A bit" off as in the single data area write example.
The second example in FIG. 6 illustrates a read operation transferring two data areas. The originator starts by sending MCB (612). The recipient responds by returning data area DATA (614). The "A bit" in this DATA frame is "ON" indicating that more data areas are to follow. After the originator receives the data area and moves it to main storage, (or elsewhere) the buffer area is free for the receipt of the next data area. The originator sends an acknowledge ACK frame (616). This ACK frame is similar to the ACK frame (606) used in the write case. The recipient responds to the ACK frame (616) by sending the next data frame DATA (618) to the originator. The "A bit" in this DATA frame is off indicating that this is the last data area. After the recipient sends the DATA frame, it sends response frame MRB (620). The "A bit" in this MRB is set to "1" since there was at least one data area preceding the response.
It should be understood that while only one operation for a single buffer set is shown in the examples of FIGS. 5 and 6, multiple buffer sets may be using the link at the same time, and that the traffic on the link consists of interleaved frames sent for the multiple buffer sets.
Returning to FIG. 3, it is noted that link control word 302 and information field 306 have independent error checking capabilities. This checking is provided by encoding of the data for serial transmission (preferably by using the 8 bit/10 bit code described in U.S. Pat. No. 4,486,739), and the CRC words 304 and 308. Transmission errors usually affect only a few bits at a time and it is unlikely that a transmission error would damage both the LC and the information field in the same frame. If the link control word is in error, the entire frame is considered lost since the receiver does not know anything about the frame such as the frame type and the buffer set number. If the link control word is not in error and only the information field is in error, the link control word provides the receiver of the damaged frame with enough information to ask the sender to retransmit the damaged frame. Since the information field is usually much longer than the LC, there is a higher chance that a transmission error will affect the information field and not the LC. This means that the most transmission errors affecting a frame can be retried using the information supplied by the link control word. This situation is illustrated in FIG. 7 which shows a write operation with an error in MCB (702). The recipient detects the error and sends a Reject (REJ) 704 frame back to the originator in which a request is made to retransmit the MCB. The originator resends MCB (706).
The state tables used for receiving frames for the originator and recipient are shown in FIGS. 8 and 12 respectively. Each buffer set maintains its own individual state. The events are listed along the top (802 and 1202 of the tables) and the originator and recipient events are described in FIGS. 10 and 14 respectively. These events are activity on the inbound link. For example, the receipt of an MRB with the "A bit" set to a "1" and no transmission errors is called MRB+A (1002). Because data area frames are much longer than MCB and MRB frames, the receipt of the link control word of these data area frames (DATA-A START 1004 and DATA+A START 1006) is an event allowing the channel to start moving the information field before all of the data area has been received thus improving performance. Listed along the left side of the state tables for FIGS. 8 and 12 are the states. The originator and recipient states for FIGS. 8 and 12 are described in FIGS. 11 and 15 respectively. For example IDLE state 1102 indicates that the originator is done with the previous operation and is ready to receive data areas and the response to the next operation. The IERR state 1104 indicates that an invalid sequence of frames was received, and the HERR state 1106 indicates that the hardware must be in error for the event to occur at this time.
Within each block 804 of the state tables there are two areas. The top of the block indicates the next state (806), and the bottom of the block indicates the action(s). The originator actions are described in FIG. 9, and the recipient action are described in FIG. 13. For example, originator action #2 (902 in FIG. 9) starts data transfer as a result of receiving the link control word of a data area.
All of the sequences shown in FIGS. 5, 6 and 7 can be traced in the state tables in FIGS. 8 and 12. For example, the operation of the recipient during the single data area write case shown in FIG. 5 can be traced in the recipient state table in FIG. 12. The recipient is normally in the IDLE state (1204) waiting for an MCB 510 frame. The arrival of this frame is the MCB+A (1206) event, and the state changes to DS where the recipient is waiting for a data area. The DATA-A START event (1216) signals beginning of DATA frame 512, and the recipient moves to the DE 1212 state waiting for the end of the data area (DE entry in FIG. 12). The DATA END event (1214) signals the successful receipt of the entire data area, and the recipient returns to the IDLE state.
As described earlier, if the link control word is in error, the entire frame is considered lost since the receiver does not know anything about the frame such as the frame type and the buffer set number. Independent timers are preferably provided for each buffer set, and these timers time from sending the MCB to the receipt of the MRB. Typically, if any of the frames is lost, the sequence does not complete and the timer times out. After a timeout, lengthy recovery procedures are invoked to determine which frame was lost and to resynchronize link operations.
Using the state tables, the originator and the recipient can detect some of these lost frames and avoid the lengthy timeout recovery procedure. Returning to FIG. 8, for example, if the originator is in the IDLE state (810) and either an MRB+A (812) (a good MRB with the "A bit" set of "1") or an MRB+A ERR (814) (An MRB in error with the "A bit" set to "1") is received, the originator assumes that a data area should have preceded the MCB. If the MRB was good, the originator sends a Reject (REJ, action 5) for the DATA frame and goes to the DRJM state (1108). In this state, the originator remembers that a good MRB was received and that it is waiting for the DATA frame to be resent from the recipient. This is the same state that the originator would be in if a DATA frame with an error in the information field was received and then the good MRB was received. If the MBR was in error, the originator sends Rejects (REJ, action 5 and 6) for both the DATA frame and the MRB and enters the DMRJ state (1110). In this state, the originator is waiting for the last DATA area and the MRB to be resent from the recipient.
The recipient also detects lost frames. Returning to FIG. 12, if the recipient is in the IDLE state (1204) and receives the start of a DATA frame (DATA-A START (1216) or DATA+A START (1210)), the recipient assumes that an MCB with the "A bit" set to "1" should have preceded the DATA frame and so it sends a Reject (REJ, action 6) for the MCB and goes to either the DEMRJ state (1502) or the DEAMRJ state (1504) depending on the value of the "A bit" in the DATA frame. In these states, the recipient remembers that a Reject was sent for the MCB and waits to receive the end of the DATA frame. When the end of the DATA frame arrives, the recipient goes to either the MRJD state (1506) or the MRJDA (1508) state depending on whether it was in the DEMRJ or DEAMRJ states. If the recipient is in the MRJD state (1506), it enters the IDLE state after the MCB is received. If the recipient is in the MRJDA state, the acknowledge (ACK) is suppressed until the MCB is received without error.
While we have described our preferred embodiments of our invention, it will be understood that those skilled in the art, both now and in the future, may make various improvements and enhancement which fall within the scope of the claims which follow. These claims should be construed to maintain the proper protection for the invention first disclosed.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US5105423 *||May 16, 1989||Apr 14, 1992||Ricoh Company, Ltd.||Digital transmission device having an error correction mode and method for shifting down a data transmission rate|
|US5168497 *||Sep 19, 1989||Dec 1, 1992||Hitachi, Ltd.||Packet communication processing method|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US5706425 *||Jan 26, 1995||Jan 6, 1998||Kabushiki Kaisha Toshiba||Timeout process circuit and receiver including this timout process circuit|
|US5717689 *||Oct 10, 1995||Feb 10, 1998||Lucent Technologies Inc.||Data link layer protocol for transport of ATM cells over a wireless link|
|US5892894 *||Dec 6, 1996||Apr 6, 1999||Nippon Telegraph & Telephone Corp.||Data re-transmission management scheme with improved communication efficiency|
|US5938786 *||Nov 30, 1995||Aug 17, 1999||International Business Machines Corporation||Simplified recovery of damaged frames in a communication link|
|US6038605 *||Oct 9, 1997||Mar 14, 2000||Emc Corporation||Method and apparatus for interfacing two remotely disposed devices coupled via a transmission medium|
|US6393595 *||Jan 15, 1999||May 21, 2002||Stmicroelectronics S.A.||Method of communication with improved acknowledgement of reception|
|US6615383 *||May 29, 1998||Sep 2, 2003||Sun Microsystems, Inc.||System and method for message transmission between network nodes connected by parallel links|
|US6647527||Mar 14, 2002||Nov 11, 2003||Stmicroelectronics S.A.||Method of communication with improved acknowledgment of reception|
|US6856795||Jan 6, 2003||Feb 15, 2005||Sharp Laboratories Of America, Inc.||Data unit detection including antenna diversity|
|US7130657 *||Oct 23, 2000||Oct 31, 2006||Lucent Technologies Inc.||Methods and systems for improving frame selection in wireless communications networks|
|US7231574 *||Sep 21, 2001||Jun 12, 2007||International Business Machines Corporation||Hardware mechanism for receiving frames from a link|
|US7324455||Mar 14, 2003||Jan 29, 2008||International Business Machines Corporation||Transfer of error-analysis and statistical data in a fibre channel input/output system|
|US7450922||Dec 7, 2004||Nov 11, 2008||Sharp Laboratories Of America, Inc.||Data unit detection including antenna diversity|
|US7562285||Jan 11, 2006||Jul 14, 2009||Rambus Inc.||Unidirectional error code transfer for a bidirectional data link|
|US7609643||Oct 9, 2007||Oct 27, 2009||International Business Machines Corporation||Transfer of error-analysis and statistical data after retry in a fibre channel input/output system|
|US7675865||Nov 15, 2007||Mar 9, 2010||International Business Machines Corporation||Transfer of error-analysis and statistical data in a fibre channel input/output system|
|US7831882||Jun 3, 2005||Nov 9, 2010||Rambus Inc.||Memory system with error detection and retry modes of operation|
|US7831888||Jun 5, 2009||Nov 9, 2010||Rambus Inc.||Unidirectional error code transfer method for a bidirectional data link|
|US7836378||Feb 21, 2008||Nov 16, 2010||Rambus Inc.||System to detect and identify errors in control information, read data and/or write data|
|US7882423||Jun 5, 2009||Feb 1, 2011||Rambus Inc.||Unidirectional error code transfer for both read and write data transmitted via bidirectional data link|
|US8132077||Jan 25, 2011||Mar 6, 2012||Rambus Inc.||Unidirectional error code transfer for both read and write data transmitted via bidirectional data link|
|US8352805||Apr 15, 2009||Jan 8, 2013||Rambus Inc.||Memory error detection|
|US8365042||Feb 16, 2012||Jan 29, 2013||Rambus Inc.||Unidirectional error code transfer for both read and write data transmitted via bidirectional data link|
|US8555116||Nov 1, 2012||Oct 8, 2013||Rambus Inc.||Memory error detection|
|US8656254||Jan 28, 2013||Feb 18, 2014||Rambus Inc.|
|US8707110 *||Sep 6, 2013||Apr 22, 2014||Rambus Inc.||Memory error detection|
|US8918703||Dec 29, 2012||Dec 23, 2014||Rambus Inc.||Memory system with error detection and retry modes of operation|
|US9031092 *||Nov 12, 2013||May 12, 2015||Pivotal Decisions Llc||Apparatus and method for transmitting LAN signals over a transport system|
|US9092352||Feb 7, 2014||Jul 28, 2015||Rambus Inc.||Memory controller with write data error detection and remediation|
|US9141479||Nov 5, 2010||Sep 22, 2015||Rambus Inc.||Memory system with error detection and retry modes of operation|
|US9170894||Mar 7, 2014||Oct 27, 2015||Rambus Inc.||Memory error detection|
|US9213591||Aug 11, 2015||Dec 15, 2015||Rambus Inc.||Electrically erasable programmable memory device that generates a cyclic redundancy check (CRC) code|
|US9262262||Sep 14, 2015||Feb 16, 2016||Rambus Inc.||Memory device with retransmission upon error|
|US9262269||Aug 26, 2015||Feb 16, 2016||Rambus Inc.||System and module comprising an electrically erasable programmable memory chip|
|US9274892||Aug 17, 2015||Mar 1, 2016||Rambus Inc.||Memory chip with error detection and retry modes of operation|
|US9298543||Jul 22, 2015||Mar 29, 2016||Rambus Inc.||Electrically erasable programmable memory device that generates error-detection information|
|US9459960||Aug 19, 2015||Oct 4, 2016||Rambus Inc.||Controller device for use with electrically erasable programmable memory chip with error detection and retry modes of operation|
|US9477547||Sep 14, 2015||Oct 25, 2016||Rambus Inc.||Controller device with retransmission upon error|
|US20030061375 *||Sep 21, 2001||Mar 27, 2003||International Business Machines Corporation||Hardware mechanism for receiving frames from a link|
|US20030100282 *||Jan 6, 2003||May 29, 2003||Srinivas Kandala||Data unit detection including antenna diversity|
|US20040193968 *||Mar 14, 2003||Sep 30, 2004||International Business Machines Corporation||Transfer of error-analysis and statistical data in a fibre channel input/output system|
|US20050096001 *||Dec 7, 2004||May 5, 2005||Srinivas Kandala||Data unit detection including antenna diversity|
|US20060277434 *||Jun 3, 2005||Dec 7, 2006||Tsern Ely K||Memory system with error detection and retry modes of operation|
|US20070162825 *||Jan 11, 2006||Jul 12, 2007||Yuanlong Wang||Unidirectional error code transfer for a bidirectional data link|
|US20070271495 *||May 18, 2006||Nov 22, 2007||Ian Shaeffer||System to detect and identify errors in control information, read data and/or write data|
|US20080034250 *||Oct 9, 2007||Feb 7, 2008||International Business Machines Corporation||Transfer of Error-Analysis and Statistical Data After Retry in a Fibre Channel Input/Output System|
|US20080063038 *||Nov 15, 2007||Mar 13, 2008||International Business Machines Corporation||Transfer of Error-Analysis and Statistical Data in a Fibre Channel Input/Output System|
|US20080163007 *||Feb 21, 2008||Jul 3, 2008||Rambus Inc.||System To Detect And Identify Errors In Control Information, Read Data And/Or Write Data|
|US20080298381 *||Aug 11, 2008||Dec 4, 2008||International Business Machines Corporation||Apparatus for queue management of a global link control byte in an input/output subsystem|
|US20090235113 *||Apr 15, 2009||Sep 17, 2009||Rambus Inc.||Memory error detection|
|US20090249139 *||Jun 5, 2009||Oct 1, 2009||Yuanlong Wang||Unidirectional Error Code Transfer for Both Read and Write Data Transmitted via Bidirectional Data Link|
|US20090249156 *||Jun 5, 2009||Oct 1, 2009||Yuanlong Wang||Unidirectional Error Code Transfer Method for a Bidirectional Data Link|
|US20110119551 *||Nov 5, 2010||May 19, 2011||Tsern Ely K||Memory System with Error Detection and Retry Modes of Operation|
|US20110209036 *||Jan 25, 2011||Aug 25, 2011||Yuanlong Yang||Unidirectional Error Code Transfer for Both Read and Write Data Transmitted via Bidirectional Data Link|
|US20140086583 *||Nov 12, 2013||Mar 27, 2014||Pivotal Decisions Llc||Apparatus and method for transmitting lan signals over a transport system|
|WO2006132840A2 *||May 26, 2006||Dec 14, 2006||Rambus Inc.||Memory system with error detection and retry modes of operation|
|WO2006132840A3 *||May 26, 2006||Apr 19, 2007||Rambus Inc||Memory system with error detection and retry modes of operation|
|U.S. Classification||714/748, 370/471, 714/749|
|International Classification||H04L1/00, G08C25/02|
|Cooperative Classification||H04L1/0083, G08C25/02|
|European Classification||G08C25/02, H04L1/00F2|
|Aug 4, 1994||AS||Assignment|
Owner name: INTERNATIONAL BUSINESS MACHINES CORPORATION, NEW Y
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:GREGG, THOMAS A.;HOKE, JOSEPH M.;ING, ALBERT;AND OTHERS;REEL/FRAME:007106/0938
Effective date: 19940804
|Aug 31, 1999||SULP||Surcharge for late payment|
|Aug 31, 1999||FPAY||Fee payment|
Year of fee payment: 4
|Aug 31, 1999||REMI||Maintenance fee reminder mailed|
|Apr 27, 2000||AS||Assignment|
Owner name: CISCO TECHNOLOGY, INC., A CORPORATION OF CALIFORNI
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:CISCO SYSTEMS, INC., A CORPORATION OF CALIFORNIA;REEL/FRAME:010756/0122
Effective date: 20000405
|Jun 14, 2000||AS||Assignment|
Owner name: CISCO SYSTEMS, INC., CALIFORNIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:INTERNATIONAL BUSINESS MACHINES CORPORATION;REEL/FRAME:010832/0878
Effective date: 20000126
|Aug 6, 2003||FPAY||Fee payment|
Year of fee payment: 8
|Jun 21, 2007||FPAY||Fee payment|
Year of fee payment: 12